axml 0.0.2
Parses elements, attributes, and text(content), and nothing more. Should be very easy to extend and modify for special cases. It is roughly as fast as libxml, although it currently reads in the entire document first (however, this is memory efficient - nodes are implemented using Struct).